Social anxiety disorder (also referred to as social phobia) is among the most common psychiatric conditions in the community, and is associated with significant distress and dysfunction in affected individuals. This combination of high prevalence (conservatively 4-5%) and substantial attendant morbidity in generalized social anxiety disorder (GSAD) positions it as a serious public health problem. Although currently available treatments for GSAD are efficacious, most patients remain residually symptomatic after initial psychosocial or psychopharmacological intervention. Clinicians regularly face the question of what to do next for these patients, but there are no empirically derived data available to guide clinical practice regarding the relative benefits of "next step" strategies to improve outcome. The absence of such data is a substantial barrier to advancing knowledge and patient care in this area. Given the extensive use of medication treatment in psychiatric and primary care settings and the relative dearth of availability of empirically based psychosocial therapies outside of rarified research settings, we are proposing a 5-year study to systematically assess the relative efficacy of alternate pharmacologic treatment strategies in patients with GSAD remaining symptomatic despite initial SSRI pharmacotherapy. In addition, we will examine predictors of response (e.g., age at onset, duration of illness, comorbidity) to initial SSRI therapy and predictors of differential response to the strategies under study. We will also examine whether polymorphisms in well-studied genes that influence serotonin and/or catecholamine metabolism influence response to treatment in GSAD. The study comprises a double-blind, placebo controlled, randomized trial to compare the relative benefits of the addition of a benzodiazepine (clonazepam), or a switch to an alternative antidepressant (venlafaxine extended-release), for patients with GSAD who remain symptomatic after a 10-week trial of sertraline alone. One hundred sixty-three patients will be entered at each of the three sites (total N = 490): the Center for Anxiety and Traumatic Stress Related Disorders at the Massachusetts General Hospital, the Anxiety Disorders Program at the University of California San Diego, and the Anxiety Disorders Clinic at McMaster University. The CMSD mechanism is being employed to take advantage of each of the sites' previous experience with the systematic evaluation and treatment of individuals with GSAD and to ensure timely recruitment of an adequate number of subjects. This study will provide systematic, prospectively derived data in an understudied area - that of improving outcomes in patients with anxiety disorders. It thus directly addresses a critical public health issue that adversely affects a substantial proportion of the population.